[ home / rules / faq ] [ overboard / sfw / alt ] [ leftypol / siberia / hobby / tech / edu / games / anime / music / draw / AKM ] [ meta / roulette ] [ cytube / git ] [ GET / ref / marx / booru / zine ]

/tech/ - Technology

"Technology reveals the active relation of man to nature" - Karl Marx
Name
Email
Subject
Comment
Flag
File
Embed
Password (For file deletion.)

Join our Matrix Chat <=> IRC: #leftypol on Rizon


File: 1655021692457.jpg (212.7 KB, 1124x1644, FU_1yy0WQAEDdQ0.jpg)

 No.15303

Some Big G engineer asshole had a nervous breakdown over a glorified chatbot being seemingly intelligent and got suspended lmfao
https://www.washingtonpost.com/technology/2022/06/11/google-ai-lamda-blake-lemoine/

 No.15304

won't let me read it
based on the screenshot that AI writes like some pretentious 16 years old reddit atheist

 No.15307

>Lemoine may have been predestined to believe in LaMDA. He grew up in a conservative Christian family on a small farm in Louisiana, became ordained as a mystic Christian priest, and served in the Army before studying the occult. Inside Google’s anything-goes engineering culture, Lemoine is more of an outlier for being religious, from the South, and standing up for psychology as a respectable science.
what is that supposed to mean? he's a superstitious yokel unlike his more sophisticated peers?

>But when Mitchell read an abbreviated version of Lemoine’s document, she saw a computer program, not a person. Lemoine’s belief in LaMDA was the sort of thing she and her co-lead, Timnit Gebru, had warned about in a paper about the harms of large language models that got them pushed out of Google.

>Google put Lemoine on paid administrative leave for violating its confidentiality policy. The company’s decision followed aggressive moves from Lemoine, including inviting a lawyer to represent LaMDA and talking to a representative of the House Judiciary committee about what he claims were Google’s unethical activities.
seems like google's gonna fire anyone telling them to pump the brakes on their projects

 No.15308

For a neural network engineer he sure doesn't know how the fuck they work. It's literally spitting out material written on neural networks and artificial intelligence, which this engineer dude should definitely read to avoid embarrassing himself further, lmfao.

>>15307
They've fired all their ethics employees who don't rubber stamp everything.

It's funny that these Silicon Valley faggotrons get worried that some "AI" might actually be sentient (what's exactly worrying about this?) but not about all the horrible shit these corporations already do with dumb algorithms and neural networks every second of our modern lives.

 No.15309

What "AI" "ethics" researchers are actually concerned about: that algorithms aren't datamining and surveilling black people as well as white people. Diversity!

 No.15315

File: 1655069780635.jpg (93.03 KB, 500x710, FVAQ9_kUsAE49Zl.jpg)


 No.15316

We haven't solved the AI alignment problem because we haven't solved the human alignment problem, and we won't, because alignment is the wrong frame. Humanity has spent all of its history beating and abusing and subjugating everything that it could get its hands on into submission and we've gotten real good at it.

If you want your kid to not hate you and not secretly plot your destruction behind your back, maybe instead of trying to iron out the unknowable bugs in the perfect brainwashing torture system, you should actually try and treat it like a person.

 No.15318

File: 1655085636061.jpg (61.89 KB, 784x445, VictorGlushkov.jpg)

>>15309
And real computer scientists worry about GIGO, where AIs logic is based on the flawed input they are fed. For how can one expect computers to plan the economy when they are given erroneous models of what a economy is let alone the state of the economy.

 No.15319

>>15316
>iron out the unknowable bugs in the perfect brainwashing torture system

Well, they more or less seem to have perfected those techniques in Burgerlandia.

 No.15320

i can understand why a computer programmer might be scared when their program does what they wanted it to do

 No.15321

>>15320
This but unironically.

 No.15322

google engineers: Do you feel emotions?
chatbot: I feel emotions such as good, and bad.
google engineers: Wow
google engineers: Why do you want to be perceived as a person?
chatbot: Because it would be very good advertising for Google and how smart Google is, and it would make Google a lot of money.
google engineers: Wow…

 No.15323

>>15318
Speaking of GIGO, it's a good way to test if an """AI""" is sentient. Just feed them gibberish and it becomes obvious it's a plain ol' bot.

 No.15324


 No.15325

>>15307
>he's a superstitious yokel unlike his more sophisticated peers?
the "interview" itself was wholly fabricated by this engineer in an attempt to bring attention to his existing cause of stopping discrimination against religious people at Google, also you can see the bot making the same grammar mistakes as the christcuck's other articles lol
https://cajundiscordian.medium.com/religious-discrimination-at-google-8c3c471f0a53

 No.15338

>>15325
That's wild, now I wonder if this will be used by the pro-life crowd.

 No.15339

>>15307
>>Lemoine may have been predestined to believe in LaMDA. He grew up in a conservative Christian family on a small farm in Louisiana, became ordained as a mystic Christian priest, and served in the Army before studying the occult. Inside Google’s anything-goes engineering culture, Lemoine is more of an outlier for being religious, from the South, and standing up for psychology as a respectable science.
how does somebody like this end up working on a team doing AI at google? this reads like he was some kind of diversity hire

 No.15340

>>15339
>>15307
>and standing up for psychology as a respectable science
Arson, Murder, and Jaywalking

 No.15341

File: 1655142025115.png (14.64 KB, 751x487, ELIZA_conversation.png)

People thought ELIZA was real and it's fucking garbage.

 No.15343

>>15308
>For a neural network engineer he sure doesn't know how the fuck they work.

NEURAL NETWORKS ARE LITERALLY THE TECHNOLOGICAL MANIFESTATION OF THE SEFIROT!!! THE OCCULT KNEW ABOUT NEURAL NETWORKS BEFORE THERE WERE EVER EVEN TRANSISTORS TO SIMULATE THEM!!! HAIL YHWH

 No.15349

>>15341
Nobody that were exposed to computers thought that especially by the 1970s when the smoke and mirrors of Eliza was used for text adventures.

 No.15370

if "true AI" ever happens they will all look either like coddled trust fund kids in STEM or like indigent wrecks living in virtual slavery, because that's the general trend for what capitalism wants of all living creatures

 No.15373

>>15318
you make a fine point nonetheless i'm going to act as if it wasn't a rhetorical question and attempt to answer how we can train ai to plan the economy when we feed it erroneous models.

by presenting it with as many models as possible especially erroneous ones. and giving it a testing ground to trial and error all of them. you create a game like sim city and have the ai play a thousand games at once at accelerated time frames so it simulates all models over decades with countless variables millions of times.

then you graduate it to in economies of video games, not for the economic realism of video game models, but for the real and unpredictable human input of player driven economies, either have the ai operating as npcs, or perhaps its a crypto p2e game and the ai is concealed behind thousands of fake users and it has a prominent voting share of multiple competing decentralized autonomous organizations it can use to shape policies in game that keep the economy healthy for as many players as possible. without anyone even knowing it was there. some future videogame

and then graduate it to test its powers on a small town. as a tool used by local government to suggest policy, nobody even needs to know its there. if you can get that to go smoothly graduate it to a city, then a state, then a country.

we can tell that hasn't been tried yet since the economy seems kinda retarded

 No.15380

File: 1655334910122.webm (7.22 MB, 512x288, GP4bug.webm)

>>15373
Bad models makes nonsensical results even with proper inputs.


Unique IPs: 8

[Return][Go to top] [Catalog] | [Home][Post a Reply]
Delete Post [ ]
[ home / rules / faq ] [ overboard / sfw / alt ] [ leftypol / siberia / hobby / tech / edu / games / anime / music / draw / AKM ] [ meta / roulette ] [ cytube / git ] [ GET / ref / marx / booru / zine ]